Metabolism of tamoxifen by isolated rat hepatocytes

Metabolism of tamoxifen by isolated rat hepatocytes

Biochemical Pharmacolog,v, Printed m Great Britain Vol. 39. No. 9. pp. 1459-1465. METABOLISM 1990 00062952/sO $3.00 + 0.00 @ 1990. Pergamon Pres...

752KB Sizes 0 Downloads 89 Views

Biochemical Pharmacolog,v, Printed m Great Britain

Vol.

39. No.

9. pp. 1459-1465.

METABOLISM

1990

00062952/sO $3.00 + 0.00 @ 1990. Pergamon Press plc

OF TAMOXIFEN BY ISOLATED RAT HEPATOCYTES

IDENTIFICATION OF THE GLUCURONIDE 4-HYDROXYTAMOXIFEN RAYMONDMCCAGUE,*~

OF

IRISB.PARR,~GUYLECLERCQ,S ON-TAI LEuNGt and MICHAEL

JARMAN? t Drug Development Section, Cancer Research Campaign Laboratory, Institute of Cancer Research, Sutton, Surrey SM2 5NG, U.K. and $ Laboratoire de CancCrologie Mammaire, Service de Mtdtcine, Institut Jules Bordet, 1000 Bruxelles, Belgium (Receiued

17 February 1989; accepted 15 December

1989)

Abstract-Metabolism of 4-hydroxytamoxifen by hepatocytes isolated from rats administered with phenobarbital and examination by TLC of the components not extractable into ethyl acetate revealed 4-hydroxytamoxifen P-glucuronide; its identity was confirmed by comparison of its ‘H NMR spectrum with that of synthetic material. This conjugate was also formed on metabolism of tamoxifen. It bound to cytosolic oestrogen receptors with only one thousandth the affinity of 4-hydroxytamoxifen and gave a correspondingly very weak inhibition of growth of the MCF-7 human breast cancer cell line. Therefore, in contrast to reported observations on the 3-glucuronide of oestradiol, the MCF-7 cells were unable to hydrolyse 4-hydroxytamoxifen glucuronide and on this evidence, formation of this metabolite is solely a deactivation pathway

Tamoxifen (1) is an excellent drug for the treatment of advanced breast cancer giving a proportion of patients a period free from disease without significant side-effects [l, 21. Its principal mode of action is thought to be as an antioestrogen, acting by competitive displacement of the tumour growth promoting hormone oestradiol from its receptor binding site [3]. Metabolism may play an important role in mediating the overall activity of this drug [4]. In particular, the metabolite 4-hydroxytamoxifen (2) [.5] has an affinity for the oestrogen receptor one hundred times that of tamoxifen and has a correspondingly greater antitumour potency than tamoxifen in vitro [6,7]. However, this high potency is not realised in viuo where tamoxifen and 4hydroxytamoxifen have comparable effectiveness [8] correlating with the low plasma levels that are achieved for the latter (cu. 200 ng/mL for tamoxifen in the sera of patients but cu. 5 ng/mL for hydroxytamoxifen) [5] and attributable to a rapid clearance 07

L NMe2

(1)

x =H

(2)x = OH * To whom correspondence

should be sent

mechanism by way of conjugation of the hydroxyl group [9]. By inference from the observation of Fromson et al. [lo] that 4-hydroxytamoxifen is liberated from faeces of tamoxifen treated patients upon treatment with /3-glucuronidase. this conjugate must be a glucuronide. We have previously shown that isolated rat hepatocytes will convert the hydroxy group of the aromatase inhibitor 4-hydroxyandrostenedione into the glucuronide [ll]. That this glucuronide is also observed as a major urinary metabolite in patients [ 121 provided evidence that rat hepatocytes provide a faithful model for metabolism in ho. Supporting this conclusion is the finding that rat hepatocytes convert tamoxifen into several of the metabolites identified in the sera of patients [13]. We now report that by examination of the fraction of polar metabolites corresponding to conjugates obtained after incubation of 4-hydroxytamoxifen or tamoxifen with isolated rat hepatocytes, the glucuronide of 4-hydroxytamoxifen can be detected directly. In addition to the presumed role of 4-hydroxytamoxifen glucuronide in the clearance of tamoxifen it was considered a possibility that the glucuronide might have significant antitumour activity in itself. Aniline mustard [NJ-di-(2-chloroethyl)aniline], like tamoxifen. undergoes metabolic hydroxylation on the aromatic ring and also a glucuronide is formed [14]. This glucuronide, which is a major serum and bile metabolite of aniline mustard [14], has been shown to be four-fold more cytotoxic than the parent drug [ 151. Although we would not anticipate that 4hydroxytamoxifen glucuronide would have a significant direct action on tumour cells, the finding that oestradiol 3-glucuronide causes a time dependent

1460

R.

MCCAGUE et

disappearance of oestrogen receptors (“processing”) in the human breast cancer cell line MCF-7 in culture, indicated that the cell line possesses glucuronidase activity liberating the potent oestrogen [16]. Moreover, oestradiol3-glucuronide partially reversed the growth inhibition of MCF-7 cell growth by 4hydroxytamoxifen despite being unable to reduce [3H]oestradiol uptake in a binding affinity assay, giving further evidence of hydrolysis by the cells. We have consequently examined the glucuronide of 4hydroxytamoxifen obtained by synthesis for its binding affinity to oestrogen receptors and for its ability to inhibit the growth of the MCF-7 cell line.

MATERIALS

AND METHODS

Materials. Z-[ring-‘4C]tamoxifen citrate (sp. act. 21 mCi/mmol and 97% chemically pure) was purchased from Amersham International, (Amersham. U.K.). Unlabelled tamoxifen, 4-hydroxytamoxifen. and methyl tetra-O-acetyl fi-D-ghicuronate were prepared by published methods [17-191. In vitro metabolism. Male adult Wistar rats fed un lib. and given phenobarbital in their drinking water (0.5 g/L) for l&14 days were used to prepare hepatocytes by a 2-step perfusion method as described previously [ll]. The freshly isolated cells (5 mL of 6-S x 10h/mL 90% viable) were incubated under an of atmosphere of O2 at 37” with a mixture [14C]tamoxifen citrate (0.8 x 10h DPM) and unlabelled tamoxifen (100 pg/mL) or with unlabelled 4-hydroxytamoxifen (100 pg/mL) in 25-mL conical flasks in a phosphate buffered saline medium at pH 7.5 [20] containing Hepes (20 mmol/L) and glucose (5 mmol/L). After incubation for 60 min the flasks were cooled in ice, the hepatocytes disintegrated by ultrasonication and the protein precipitated with 4 volumes of acetone. The acetone was removed under vacuum and after adjustment of the pH to 9.5. non-utilized substrate and phase I nonpolar metabohtes were extracted with ethyl acetate. Conjugates remaining in the aqueous fraction were concentrated on Sep-pak cartridges, eluted with methanol and then subjected to reverse-phase chromatography on Whatman KC 18F plates (20 x 20 cm). eluting with methanol : water: 40% tetrabutylammonium hydroxide (80 : 20 : 1) and using synthetic 4-hydroxytamoxifen glucuronide as a marker to locate the position of the metabolite glucuronide. In case of [‘“Cltamoxifen a Berthold automatic analyser was used to locate the radioactive peaks and to determine the proportion of label in the peak corresponding to 4-hydroxytamoxifen glucuronide. NMR spectra. Proton NMR spectra (250MHz; internal tetramethylsilane reference) were recorded on a Bruker AC250 spectrometer. Solutions of metabolites in methanol extracted from the chromatography plates were filtered, concentrated to dryness under reduced pressure, last traces of solvent being removed at 0.01 mm Hg over 1 hr. The resulting residues were redissolved in the appropriate deuterated solvent (0.5 mL) in 5 mm NMR tubes. For metabolites, data was accumulated over 1 hr (1500 scans).

al.

Binding of the glucuronide (4a) to oestrogen receptors [21]. Calf uterine cytosol was incubated at 18” for 30 min with 5 x 1O-y M [“Hloestradiol in the absence and presence of increasing amounts ( 1Omy to 10m5M) of the glucuronide (4a) or unlabelled oestradiol (control). Unbound compounds were then removed using dextran-coated charcoal and the amounts of oestrogen receptor bound [‘Hloestradiol were measured. The relative concentration of oestradiol and test compound required to achieve 50% inhibition of [“Hloestradiol binding is the relative binding affinity (RBA), i.e. RBA = ([I,,,]oestradiol/[I,,,] test compound) x 100. Effect of the glucuronide (4a) on MCF-7 cell growth. The actions of 4-hydroxytamoxifen glucuronide, 4hydroxytamoxifen and tamoxifen on MCF-7 cell growth were compared by measuring the amounts of DNA after 5 days of culture according to the protocol described previously [22]. Action of the compounds on ER levels was estimated after 3 days of culture when amounts of receptor were measured in cytosol extracts by multipoint DCC assay using [3H]oestradiol as labelled ligand according to the accepted method 1231. SYNTHESIS

The glucuronide of 4-hydroxytamoxifen was synthesized as shown in Scheme 1; the method being essentially that used by Hadd et u[. [24]. for the synthesis of the anomeric 17-glucuronides of ethynyloestradiol. Thus, treatment of 4hydroxytamoxifen as a mixture of trans (2a) and cis (2b)-isomers

with

methyl

tetra-Q-aCetyi-B-D-glU-

curonate in the presence of excess tin (IV) chloride as Lewis acid gave the glucuronate ethers (3a, 3b). There was no advantage in using the pure trunsgeometric isomer of 4-hydroxytamoxifen since rapid isomerization to a &-tram mixture was catalysed by the tin chloride. The condensation gave high selectivity for the required /Ganomer but with extended reaction time there was isomerization to the acanomer which proved difficult to remove. It was consequently preferable to work up the reaction when it was only about two-thirds complete. The tram- and cis-b-glucuronates were separated by chromatography on silica and were individually hydrolysed to the potassium salts of the glucuronides (4a) and (4b). As confirmation of structure, these trans- and cis-glucuronides were treated with @glucuronidase (from Escherichiu co/i) in phosphate buffer when there was, in each case, a rapid hydrolysis to give the corresponding isomer of 4-hydroxytamoxifen (identified by reverse phase TLC). I-O-[4-[l-[4-[2-(dimethyluminojethoxyjMethyl phenyl]-2-phenyl-(Z and E)but-l-enyl[phenyl]-2,3.4tri-0-ucety/-P-D-glucuronute. To a stirred solution of methyl-l ,2.3,4-tetra-0 acetyl-p-D-ghicuronate (1.72 g, 4.58 mmol) in dichloromethane (10 mL) at 20” was added tin (IV) chloride (1.6 mL, 13.6mmol). When the solution became mauve (ca. 20 min) a suspension of 4hydroxytamoxifen (1 : 1 mixture of Z and E isomers; 1.17 g, 3.03 mmol) in dichloromethane (10 mL) was added. After 24 hr (TLC indicated reaction about

Metabolism of tamoxifen by rat hepatocytes

ea.

1461

7-b) (3a, 3b)

(Z&u)

R’=Ph,

R2=Et (tram)

(Zb-4b) R’=Et, R’=Ph (cis)

‘OH (4a, 4b)

Scheme

1, Synthesis

of the glucuronide

two-thirds complete), the mixture was poured into saturated aqueous sodium hydrogen carbonate (125 mL) and the products extracted with ether (2 x 100 mL). The combined extracts were dried with sodium sulphate and concentrated. The residue was applied to a column of silica gel (Merck 15111; 40 g). Elution with dichloromethane-methanol (50: 1) gave the methyl triacetyl+glucuronate as a 3:4 mixture of E-(trans) and Z-(cis) isomers (453 mg, 21%). Separate isomers were obtained by further chromatography on silica but eluting with ether-triethylamine (100 : 1) giving (i) the E-(trans)isomer (3a) (44mg) as a powder, m.p. 120-122” (from light petroleum, b.p. SO-100°), NMR (CDCQ: 60.91 (t, J 7.4H2, 3H, W&H,), 2.04, 2.05, 2.07 (3 x s, 9H, 3 x COCH3), 2.31 (s, 6H, NMe2), 2.45 (q, J 7.4 Hz, 2H, CHQf,), 2.68 (t, J 5.8 Hz, 2H, OCH,CH,N), 3.73 (s, 3H, CO*Me), 3.95 (t, J5.8 Hz, 2H, 0CH2CH2N), 4.19 (m, lH, H5),5.17(d,J7.1Hz,lH,H-l),5.25-5.38(m,3H,H6.54 (d, J8.7 Hz, 2H, ArH ortho to 2,3,4), 0CH2CH2N), 6.74 (d, J8.7 Hz, 2H, ArH ortho to 0CH2CH2N), 6.96 (d, .I 8.6 Hz, 2H, ArH), 7.077.18 (m, 7H, ArH); Anal. Calc. for C39H45011:C, 66.6; H, 6.4; N, 2.0. Found: C, 66.3; H, 6.5; N, 2.0% and (ii) the Z-(cis)-isomer (3b) (170mg) as a powder, m.p. 144-146” (from light petroleum, b.p. SO-loo’), NMR (CDC&): 6 0.92 (t, J7.4Hz, 3H, CH,CH,), 2.00, 2.02, 2.02 (9H, 3 x s, 3 x COCH,). 2.39 (s, 6H, NMe2), 2.46 (q, J7.4 Hz, 2H, CH,CH,), 2.79 (t, J5.7 Hz, 2H, OCH,CH,N), 3.67 (s, 3H, 4.07-4.14 (m, 3H, 0CH7CH2N and C&Me), H-5) 5.00 (d, J7.0 Hz, lH, H-l), 5.18-5.31 (m, 3H, H-2,3,4), 6.62 (d,J8.8Hz, ArH), 6.78 (d, 8.8 Hz, ArH), 6.89 (d, J 8.7 Hz, ArH ortho to OCH2CH2N), 7.06-7.25 (m, 7H, ArH); Anal. Calc.

of 4-hydroxytamoxifen

for C39Hd5011: C, 66.6; H, 6.4; N, 2.0. Found: 66.5; H, 6.5; N, 2.0%.

C,

l-0-[4-[1-[4-[2(Dimethylamino)erhoxylphenyl]-2phenyfbut-1-enyfbheny[l-P-D-glucuronic acid (4hydroxytamoxifen glucuronide) potassium salt Trans-(E)-isomer (4a). The trans-(Qglucuronate (3a) (16.5 mg, 0.02 mmol) was added to a solution of 85% potassium hydroxyde (13.3 mg, 0.20 mmol) in methanol (0.25 mL). After 2 hr, ethyl acetate (0.5 mL) was added and the mixture concentrated to dryness. The residue was applied to a column of silica gel (Merck 15111; lg). Elution with chloroformmethanol (3: 1) gave the hydrated potassium salt of trans-4-hydroxytamoxifen glucuronide as a glass (12.7 mg, 87%), NMR (acetic acid-d4): 60.92 (t, J 7.4Hz, 3H, CH,CH,), 2.49 (q, J 7.4Hz, 2H, CH,CH*), 2.97 (s, 6H, NMe2), 3.55 (br.t., J 6Hz, 2H, OCH,CH,N), 3.78-3.92 (m, 3H, H-2,3,4), 4.194.28 (m, 3H, H-5 and 0CH2CHZN), 5.18 (d,J7.2 Hz, lH, H-l), 6.61 (d, J 8.7 Hz, 2H, ArH ortho to 0CH2CH2N), 6.83 (d, J 8.7Hz, 2H, ArH meta to 0CH2CH2N), 7.OG7.20 (m, 9H, ArH); Anal. Calc. for C,,H,,KN0,.11H20 C, 61.1; H, 6.25; N, 2.2. Found: C, 60.9; H, 6.3; N, 2.3%. Cis-(Z)-isomer (4b). Similar treatment of the cis(Z)-glucuronate (62.5 mg, 0.089 mmol) with 85% potassium hydroxide (53 mg, 0.80 mmol) in methanol (1 mL) by the method described for the transisomer gave the potassium salt of cis-4-hydroxytamoxifen glucuronide as a glass (48.8mg, 91%); NMR (acetic acid-d4): 6 0.93 (t, J 7.4Hz, 3H, CH,CH,), 2.49 (q, J 7.4 Hz, 2H, CH,CH*), 3.04 (s, 6H, NMe), 3.65 (br.t, .I 6 Hz, 2H, OCH,CH,N), 3.68-3.85 (m, 3H, H-2,3,4), 4.11 (d, lH, J 9.1 Hz, H-5), 4.39 (br.t., J 6 Hz, 2H, OCH2CH2N), 5.01 (d,

R. MCCAGUEet al.

1462

J 7.5 Hz, lH, H-l), 6.73 (d, J 8.8 Hz, 2H, ArH), 6.82(d,J8.8Hz,2H,ArH),6.96(d,J8.7Hz,ArH ortho to OCH,CHIN), 7.09-7.22 (m, 7H, ArH); Anal. Calc. for C,,H,,KNO,: C, 63.9; H, 6.0; N, 2.3%. Found: C, 64.2; H. 6.4; N 2.4%. Hydrolysis of the fi-glucuronides

by glucuronidase

To a solution of the glucuronide (4a; 1.2 mg) in pH 6.3 phosphate buffer (0.5 mL) at 20” was added a solution of /3-glucuronidase (Escherichia coli K12; 200 p/mL; 0.1 mL). Within 30 set the mixture had become cloudy owing to deposition of 4hydroxytamoxifen and after 2 min the hydrolysis was complete, and reverse-phase TLC showed only the presence of trans-(Z)-4-hydroxytamoxifen which could be isolated by ether extraction. Similar treatment of the glucuronide (4b) liberated cis-(E)-4hydroxytamoxifen. No hydrolysis of (4a) or (4b) took place in the presence of buffer alone. RESULTS Identification of 4-hydroxytamoxifen

glucuronide

The synthetic glucuronides could be observed on reverse-phase TLC after elution with a solvent system of 80 : 20 : 1 methanol : water: 40% aqueous tetrabutylammonium hydroxide. The addition of the tetraalkylammonium salt gave rise to improved resolution relative to the more routinely used diethytetrabutylammonium hydrogen lamine. Indeed, sulphate has been used in the reverse-phase TLC analysis of bilirubin glucuronides [25]. There was complete separation of the trans’ and cis isomers in this system, the trans isomer being the more mobile (tram isomer, Rf = 0.55; cis isomer Rr = 0.49). Initially hepatocyte metabolism of tram-6 hydroxytamoxifen rather than of tamoxifen was studied in order to maximize the production of glucuronide. The fraction of metabolites not extractable into ethyl acetate and corresponding to conjugates was examined on RPTLC revealing by comparison with the synthetic materials, the trans-glucuronide as the major metabolite together with a small amount of its cis-isomer. The formation of some cis-glucuronide is not surprising; it will have formed by isomerization of the 4-hydroxytamoxifen prior to the conjugate formation, a process which is facile owing to conjugation between the hydroxyl group and central olefinic bond and which has been shown to take place under cell culture conditions [26]. The metabolites could be visualized by UV light by virtue of the triarylethylene chromophore or by treatment with a naphthoresorcinol spray followed by heating when a mauve coloration indicated the presence of the glucuronide moiety [27]. In attempting to confirm the identity of the metabolites mass spectroscopy was initially employed. This technique had been used successfully to identify the glucuronide of 4-hydroxyandrostenedione as its permethylated derivative 1121. However, such * Trans and cis refer to the relationship of the ethyl group and ring bearing the basic side-chain on the olefinic linkage. For 4-hydroxytamoxifen, the Z-isomer is frans but for its glucuronide

the Z-isomer

is cis.

I

k A

a

1?1-1-C

hLL D

ALE I

h_u F

T

61.2

0

1

6.8

8

I

6.4

Fig. 1. Part of the proton NMR spectra of2 and E isomers of 4-hydroxytamoxifen and @glucuronides recorded at 250 MHz in methanol-d,. Individual spectra are (A) trans(E)-glucuronide-synthetic; fans-( Q-glucuronide(B) metabolite; (C) civ-(Qglucuronide-synthetic; (D) cis-(Z)glucuronide-metabolite; (E) Ircrns-(Z)-4-hydroxytamoxifen; (F) ci.v-(E)+hydroxytamoxifen.

permethylation of 4-hydroxytamoxifen glucuronide would cause quaternization of the dimethylamino function, and the resultant salt would lack the necessary volatility. A methane chemical ionization (CI) mass spectrum was attainable on the synthetic glucuronide when in the form of the tetrabutylammonium salt. In the CI spectrum, an ion at m/z = 619 (0.9% of base peak) corresponds to the butyl ester of the glucuronide (c.f. Ref. 28). Unfortunately, mass spectrometry did not prove sufficiently sensitive in this case since we were not able to obtain a spectrum on the metabolite [29]. Confirmation of structure was made by proton NMR spectroscopy on both the tram and cis glucuronide metabolites. Using d,-methanol as the solvent spectra of the metabolites were obtained and matched the spectra of synthetic material. The glucuronides showed characteristic patterns in the region of the spectra containing the aromatic proton resonances (Fig. 1). The electron withdrawing influence of the glucuronide moiety causes the AB quartet signal seen for the hydroxylated phenyl ring to reduce to a singlet. In the cis-glucuronide this singlet is at

Metabolism of tamoxifen by rat hepatocytes

6

i

Fig. 2. Full proton

6

j

;

j

1463

i

i

NMR spectra of trans-4-hydroxytamoxifen glucuronide recorded at 250 MHz in acetic acid-d,. (A) Synthetic material; (B) metabolite.

lower frequency than the other aromatic signals owing to the through space shielding influence of the adjacent phenyl ring, whereas in the trans-glucuronide, the singlet is conversely in a high frequency position. An important feature of the NMR spectra of the glucuronide is the signal for the anomeric proton on the sugar ring for this confirms the presence of glucuronide ether linkage and reveals the stereochemistry. This signal was buried in d4-methanol under the signal for exchangeable protons. The resonance is however seen when d,-acetic acid is the solvent (Fig. 2, spectrum A) as a doublet at 65.19. The coupling constant of 7.2 Hz is characteristic of a fi-glucuronide, an a-glucuronide would give a narrower doublet J = 3.6 Hz [24]. The trans-glucuronide metabolite (but not the &-isomer) was contaminated by an unknown component that could not be separated by chromatography and was present in each of three occasions that the metabolite was isolated. This contaminant was also a P-glucuronide since it gave a doublet at 6 5.12, J = 7 Hz (Fig. 2, spectrum B). When tamoxifen was the substrate for the hepatocyte metabolism, the glucuronide of 4-hydroxytamoxifen was again observed but only in small quantity. The presence of the radioactive label allowed quantification. The conjugate fraction comprised 16% of the added tamoxifen, and following reverse-phase TLC separation, the 4-hydroxytamoxifen glucuronide, identified by comparison of its chromatographic mobility with a marker of synthetic material, amounted to 23% of the conjugate fraction, a yield of 3.5% from tamoxifen. These figures compare favourably with results from our previous study [13], where the bulk glucuronide fraction was hydrolysed and analysed, and when 4hydroxytamoxifen was recovered from 20% of the conjugate fraction in a yield of 1.5% from tamoxifen. In addition to 4-hydroxytamoxifen glucuronide, we obtained several bands of lower mobility on RPTLC. By reference to our previous study [13] these will be glucuronides formed from metabolites with the side

chain demethylated, N-oxidized, or modified to lhydroxyethyl. The identity of the band attributed to 4-hydroxytamoxifen glucuronide was confirmed by @-glucuronidase hydrolysis and repeat RPTLC analysis whereupon the mobility matched that of a 4hydroxytamoxifen marker. Endocrinological properties of tamoxifen /3-glucuronide in vitro

trans-4-hydroxy-

The synthesized glucuronide (4a) had an affinity for cytosolic oestrogen receptors (RBA = 0.1, oestradiol reference = 100) that was lower than that of tamoxifen (RBA = 1) and only one thousandth that of 4-hydroxytamoxifen (RBA = 100). Its antitumour potency was determined by measuring its inhibition of the rate of growth of MCF-7 human breast cancer cells in culture. In the absence of oestradiol the glucuronide (4a) inhibited growth only at the highest concentration used (10m6 M) (Table 1). It is therefore about ten-fold weaker an inhibitor than tamoxifen which gave inhibition at 10 -‘M. This poor antitumour effect of the glucuronide therefore corresponds to its RBA value. There can be no significant degree of hydrolysis of the glucuronide to 4-hydroxytamoxifen during the 4 days of the assay since 4-hydroxytamoxifen is a potent antioestrogen and even 1% hydrolysis (formation of 10eXM 4hydroxytamoxifen from 10mh M glucuronide) would have produced a more marked growth inhibition. In the presence of lo-’ M oestradiol, the glucuronide was ineffective as an inhibitor of cell growth at 10m6M. The view that the glucuronide is not hydrolysed is supported by measuring remaining oestrogen receptor levels in cytosol extracts after 3 days of culture in the presence of the compounds. Thus in this measure of “oestrogen receptor processing” [30.31], 4-hydroxytamoxifen produced a 50% reduction at a concentration between lo-“‘M and 10m9M, and a total disappearance of receptor at lo-* M. On the contrary, the glucuronide was very much less effective giving a 50% reduction between lO_‘M and

1464

R. MCCAGUE et al. Table

1. Effect

on growth

of the MCF-7

Amount Compound Tamoxifen (1) 4-Hydroxytamoxifen Glucuronide (4a)$ * t $ 35.2 37.7 § The qP

(2)

of DNA

cell line (pg ADN)

lO_ ‘M

10 I’M

12.4 t 0.6 (105)14.8 -c 1.4 (41)l 12.6 + 1.5 (107)

7.6 f 1.3 (64)1/ 4.s * 0.9 (3X)1 12.4 ? 1.6 (105)

f SD* IOmhM 4.7? 1.0 (4O)iil 4.9 -t 0.5 (42)p 7.2 t 1.2 (hl)fi

Each result is the average of four determinations. Percentage of control value, the amount of DNA in control was I I .8 -+ 0.7 /Lg ADN. In the presence of lO_*M oestradiol. 10mhM glucuronide gave an amount of DNA = + 2.2 pg (196%) whereas tamoxifen gave 26.3 + 2.8 pg (147%) and ocstradiol alone gave + 3.2 pug (210%). One way variance analysis showed that the compounds produce a significant effect on growth; Newman-Keuls test showed that they were effective at either 10 ‘. lO_‘or 10 ’ M (11P < 0.05. < 0.01).

10mhM which is the same range as for tamoxifen (control oestradiol: 50% between lo-” M and 10-“M). This behaviour of the glucuronide is in contrast to the findings with oestradiol3-glucuronide

(161. DISCUSSION

We have demonstrated that the /3-glucuronide of 4-hydroxytamoxifen is readily observed on reversephase TLC after incubation of 4-hydroxytamoxifen or tamoxifen by rat hepatocytes. Further to our lack of success in obtaining a mass spectrum of the metabolite owing to incompatability of the basic sidechain with derivatization procedures, but the ease of obtaining an NMR spectrum, we recommend the greater use of NMR spectroscopy for the identification of glucuronide metabolites in general. The studies on MCF-7 cell growth inhibition lead to the deduction that MCF-7 cells cannot hydrolyse the Pglucuronide of 4-hydroxytamoxifen. This finding is in marked contrast to an earlier study which provided evidence that the 3-glucuronide of oestradiol was hydrolysed [16]. Yet, the P-glucuronidase derived from E. co/i was capable of hydrolysing the 4-hydroxytamoxifen glucuronide. It would thus seem that @-glucuronidase in MCF-7 cells has specificity for the glucuronide of the natural hormone and maybe this is also the situation in normal breast tissue, providing a mechanism for control of the concentration of hormone in the breast in addition to the desulphation of circulating oestrogen sulphates

1321. In the absence of any hydrolysis by the tumour cells, if it assumed that MCF-7 cells are representative of breast tumours in general, then the formation of the glucuronide of 4-hydroxytamoxifen by the liver can solely represent a deactivation pathway and fully explains the somewhat poor activity of 4-hydroxytamoxifen in Go. A similar situation inhibitor 4-hydroxyaromatase affects the androstenedione, and presumably there is a similar problem for other drugs that contain hydroxyl groups where long duration of action is desired, as it is in cancer chemotherapy by continuous endocrine manipulation. Consequently approaches aimed at circumventing the glucuronidation process should

be worthwhile in the development of new drugs of increased in uiuo potency. Three strategies are: (i) addition of a second agent to divert the glucuronidation process. as in the use of sodium valproate to retard conjugation of 4-hydroxyandrostenedione [33]; (ii) structural modifications to impede glucuronidation, possibly by placement of an appropriate function close to the hydroxyl group [29]; and (iii) replacement of the hydroxyl group by a function that lacks the propensity for glucuronidation but possesses the binding capacity of the hydroxyl group or mimics its role in a mechanism based inhibition process. For the purpose of estimating the are presently hepatocyte

determining come.

effectiveness of these strategies, under continuing investigation,

which the rat

model should prove a valuable tool in whether unwanted conjugation is over-

Acknowledgements-This investigation was supported by grants from the Cancer Research Campaign and Medical Research Council to the Institute of Cancer Research, Royal Cancer Hospital and from the Fonds Cancerologique de la Caisse G&&ale d’Epargne et de Retraite de Belgique to the Institut Jules Bordet. We thank Mr M. H. Baker for the mass spectra. One of us (O.-T. L.) thanks the Cancer Research Campaign for a Research Studentship.

REFERENCES 1. Heel RC, Brogden RN. Speight TM and Avery GS, Tamoxifen: a review of its pharmacological properties and therapeutic USC in the treatment of breast cancer. Drugs 16: l-24, 1978. 2. Furr BJA and Jordan VC. The pharmacology and clinical uses of tamoxifen. fharmacol Ther 25: 127-205. 1984. 3. Nicholson RI. Daniel P. Gaskell SJ. Syne JS. Davies P and Griffiths K. The biochemistry of tamoxifen therapy of breast cancer. Antihormones 253-267. 1979. 4. Jordan VC, Metabolites of tamoxifen in animals and man: identification. pharmacology and significance. Breast Cancer Res Treat 2: 123-138, 1982. 5. Daniel CP, Gaskell SJ. Bishop H and Nicholson RI. Determination of tamoxifen and metabolites in patients with advanced breast cancer using gas chromatographymass spectrometry. J Endocrinol83: 401-408. 1979. 6. Jordan VC, Collins M. Rowsby L and Prestwich G. A monohydroxylated metabolite of tamoxifenwith potent

Metabolism

7.

8.

9

10.

Il.

12.

13.

14

15.

16.

17.

18.

19.

20

of tamoxifen

antioestrogenic activity. J Endocrinol 75: 305-316, 1977. Binart N, Catelli MG, Geynet C, Puri V. Hahnel R. Mester J and Baulie EE, monohydroxytamoxifen: an antioestrogen with a high affinity for the chick oviduct oestrogen receptor. Biochem Biophys Res Commun 91: 812-818, 1979. Wakeling AE and Slater SR. Estrogen receptor binding and biological activity of tamoxifen and its metabolites. Cancer Treat Rep 64: 741-744. 1980. Jordan VC and Naylor KE, The antitumour activity of tamoxifen and monohydroxytamoxifen. A comparative study in the rat. Proc Br Pharmacol Sot 64: 376p-377p, 197X. Fromson JM. Pearson Sand Bramah S. The metabolism of tamoxifen (ICI 46. 474). Part II: in female patients. Xenobiotica 3: 711-714. 1973. Foster AB, Jarman M, Mann J and Parr IB. Metabolism of 4-hydroxyandrost-4-ene-3,17-dione by rat hepatocytes. .I Steroid Biochem 24: 607-617. 19X6. Goss PE. Jarman M. Wilkinson JR and Coombes RC. Metabolism of the aromatase inhibitor 4-hydroxyandrostenedione in oioo. Identification of the glucuronide as a major urinary metaholitc in patients and biliary metabolite in the rat. J Steroid Biochem 24: 619622, 1986. Parr IB, McCaguc R. Lcclercq G and Stoessel S. Metabolism of tamoxifen by isolated rat hepatocytes: Identification of I-[4-(2-hydroxyethoxy)phcnyl]-l-(4hydroxyphenyl-2-phenyl-I-hutene and the dependence of N-oxidation on oxygen availability. Biochem Pharmacol36: 1513-1519, 1987. Connors TA, Farmer PB. Foster AB, Gilsenan AM, Jarman M and Tisdale MJ. Metabolism of aniline mustard [N,N-di-(2-chloroethyl)aniline]. Biochem Pharmacol22: 1971-1980, 1973. Bukhari MA, Everett JL and Ross WCJ. Aryl-2-halogenoalkylamines-XXVI. Glucuronic, sulphuric and phosphoric esters of p-di-2-chloroethylaminophenol. Biochem Pharmacol21: 963-967, 1972. Devleeschouwer N. Legros N, Olea-Serrano N, Paridaens R and Leclercq G. Estrogen conjugates and serum factors mediating the estrogenic trophic effect of MCF-7 cell growth. Carlcer Res 47: 58X3-5887, 19X7. McCague R. Stereoselective olefin formation from the dehydration of I-(p-alkoxyphenyl)-1.2.diphenylbutan1-01s: Application to the synthesis of tamoxifen. J Chem Sot Perkin Tram 1: lOll-1015. 1987. McCague R. The ubc of the perfluorotolyl protecting group in the synthesis of pure Z- and E-isomers of 4hydroxytamoxifen. J Chem Res Synop 5X-59. 1986. miniprint 771-793. 19X6. Bollenhack GN, Long JW, Benjamin DG and Lindquist JA. The synthesis of aryl-o-glucopyranosiduronic acids. J Am Chem Sot 77: [email protected], 1955. Aune H. Hals PA, Hansen BI and Arbakke J. Effect

by rat hepatocytes

21

22

23

24

25

26

27

2x

29 30.

31.

32.

33.

1465

of diethyl ether on the formation of paracetamol sulfate and glucuronide in isolated rat hepatocytes. Pharmacology 28: 67-73, 1984. Leclercq G, Dehoel MC and Heuson JC, Affinity of estradiol mustard for estrogen receptors and its enzymatic degradation in uterine and breast cancer cytosols. Int J Cancer 18: 75C-756, 1986. Leclercq G, Devleeschouwer N and Heuson JC, Guide lines in the design of new antiestrogens and cytotoxiclinked estrogens for the treatment of breast cancer. J Steroid Biochem 19: 75-85, 1983. EORTC Breast Cancer Co-operative Group, Revision of the standards for the assessment of hormone receptors in human breast cancer. Eur J Cancer 16: 151s 1515, 1980. Hadd HE, Slikker Jr W, Miller DW, Helton ED, Duax WL. Strong PD and Swenson DC, Synthesis and characterisation of the anomeric pair of 17pglucuronides of ethynyloestradiol. J Steroid Biochem 18: X1-87. 19x3. Jansen PLM. /$Glucuronidase-resistant bilirubin glucuronide isomers in cholestatic liver disease-determination of bilirubin metaholites in serum by means of high pressure liquid chromatography. Clin Chim Acta 110: 309-317, 1981. Katzencllenbogen JA, Carlson KE and Katzenellenhogen BS, Facile geometric isomerization of phenolic non-steroidal estrogens and antiestrogens: limitations to the interpretation of experiments characterising the activity of individual isomers. J Steroid Biochem 22: 589-596, 1985. Parikh I. MacGlashar DW and Fenselau C, Immobilised glucuronyltransferase for the synthesis of conjugates. J Med Chem 19: 29&299. 1976. Emary WB, Shen Z. Cooks RG, Plakunov I and Baird WM, Nucleophilic reactions between biological conjugates and tetraalkylammonium ion pair reagents during desorption chemical ionisation. Biomed Environ Mass Spectrom 15: 571-576, 1988. Leung O-T, Synthesis of analogues of tamoxifen, PhD Thesis, University of London, 1985. Horwitz KB and McGuire WL. Nuclear mechanism of estrogen action. Effect of estradiol and antiestrogens on estrogen receptors and nuclear receptor processing. J Biol &em 25j: X185-8191. 1978. . . Gvling M and Leclerca G. Estrogen and antiestroeen interaction with estrogen receptor of MCF-7 cell& relationship between processing and estrogenicity. J Steroid Biochem 29: l-8. 1988. Santner SJ. Feil PD and Santen RJ, In situ estrogen production via the estrone sulfatase pathway in breast tumors: relative importance oersus aromatase pathway. J Clin Endocrinol Metab 59: 29-33. 1984. Parr IB, Rowlands MG, Houghton J and Jarman M, Inhibition of the formation of 4-hydroxyandrostenedione glucuronide by valproate. Biochem Pharmacol37: 4581-4583, 1988.